Plenary Talks
| Plenary Talk I | Smart Sensing on the Safety of Underground Works |
| Date / Time | 1 June 2018, Friday / 13:30 - 15:00 hrs |
| Speaker | Hongwei Huang, Qingtong Li and Dongming Zhang Key Laboratory of Geotechnical and Underground Engineering and Department of Geotechnical Engineering, College of Civil Engineering, Tongji University |
| Chairs | Hongwei Huang, Limin Zhang |
It is widely accepted that the fast inspection and timely monitoring on structural performance of tunnels are of great importance to the structural and operational safety. However, as the number of constructed tunnels increases in an incredible speed, the conventional inspection and monitoring method can hardly fulfill the requirement due to the intensive labor work and low efficiency of data processing. On the other hand, the recent developing artificial intelligence (AI) based technique has the great advances in the pattern recognition and data analysis. In this paper, an AI-based smart sensing framework is proposed for fast detection, accurate identification and real-time monitoring of structural defects of shield tunnel of Shanghai metro. First, the general defect images for the inner surface of tunnel are captured by the self-developed image acquisition equipment, named as Moving Tunnel Inspection (MTI-200a) system. By using the deep learning (DL) with fully convolutional network (FCN), the hidden defects are quickly localized and characterized. Then, a wireless sensing network (WSN) with micro-electronic mechanical system (MEMS) based tilt node is applied for the real-time monitoring of tunnel convergence at the position where the significant defects are located. The evolution of tunnel deformation can be tracked with the WSN system. Thus, the defects can be sensed in a smart way along the lifetime of tunnel operation. The smart sensing framework has been successfully applied into the Shanghai metro tunnels for validation. The defects discussed in this paper include the convergence of tunnel horizontal diameter and structural defects, such as cracks and water leakage. The great superiority of AI-based methods comparing with frequently-used conventional methods should be helpful to the preventive maintenance of operational tunnels.
Keywords: Artificial intelligence, Smart sensing, Metro tunnel, Convergence deformation, Structural defect
| Biography |
Hongwei Huang is a "Yangtze River Scholar" Distinguished Professor and is presently the Dean of Graduate School of Tongji University . Hongwei is a member of Geotechnical Risk Management Committee of the ASCE, work group of International Tunnelling Association (ITA), International Society of Rock Mechanics (ISRM), International Association for Engineering Geology and the Environment (IAEG), International Society for Soil Mechanics and Geotechnical Engineering (ISSMEGE), and President of Engineering Risk and Insurance Research Branch in Chinese Society of Civil Engineering, Special Committee of Risk Management of Tunnel and Underground Works Branch of China Civil Engineering Society. His research activity focuses on Geotechnical and Tunnel Engineering Risk Assessment, Risk Warning and Risk Control on Tunnel and Underground Structure Health Monitoring. He serves as an Editorial Member of TUST, Georisk and 5 other Chinese Journals.
His research activity focuses on Geotechnical and Tunnel Engineering Risk Assessment, Risk Warning and Risk Control Tunnel and Underground Structure Health Monitoring and Testing. He serves as an Editorial Member of Georisk and 5 other Chinese journals.
